• Total Users: Unique users who asked a question and received an answer
• % of Active System Users: Users who asked a question ÷ active users for this system
• Total Questions: Total answered questions, excluding those with no user response
• Avg. Questions per User: Total questions ÷ unique users
• Total Unanswered Questions: Questions AI Answers couldn't respond to — due to no matching knowledge source or an unclear query
• % Unanswered Rate: Total unanswered questions ÷ total answers

• Users vs. Questions (trend graph): Tracks total questions and unique users over time. Hover over a date for exact counts.
• Most Popular Topics (word cloud): AI-generated topics from all user chats. Larger, darker topics have the most questions. Hover over a topic for question count and popularity ranking.
• Topics Overview (table): Per-topic breakdown of users, questions, failed answers, and positive/negative feedback. Click a topic's user count for a deep-dive pie chart segmented by user attribute.

• High unanswered rate → review and expand knowledge sources in the AI Center for the most common unanswered topics
• Topics with high failed answers → add or update Custom Q&A entries for those topics
• High negative feedback on a topic → review the knowledge source content and prompt accuracy for that area
• Low total users relative to active system users → verify AI Answers is enabled and surfaced in the Pro Menu for this system

(Coming Soon!)

• Active Users: Users who visited the system while WalkMe was loaded, regardless of interacting with a WalkMe item
• Users Exposed: Unique users who were shown the Action Bar widget on screen
• Users Interacted: Unique users who clicked a launcher or performed an action in the Action Bar
• % of Active System Users: Users who saw the Action Bar ÷ total active users
• % of Users WalkMe Was Available To: Users who used an Action Bar action ÷ users who had the Action Bar available to them
• Output Views: Total outputs fully displayed from a launcher (output must be fully displayed to count)
• Copy/Insertion Times: Number of times a generated output was copied or inserted by the user
• Positive / Negative Feedback: User feedback clicks per launcher

(Coming Soon!)

• User Analysis Funnel: Active users → Users Exposed → Users Interacted. Shows the drop-off at each stage and calculates the percentage ratios.
• User Adoption Trend: Time-series view of the funnel — useful for tracking impact of new launcher deployments or user communications.
• Launchers Overview (table): All Action Bar launchers with status, total users, total clicks, output views, copy/insertion times, and feedback. Two views:
  • Builder Created — out-of-the-box and console-built launchers, including quick actions and writing/reading tools
  • End-User Created — launchers created by end-users via the custom actions feature in the Action Bar
• Launcher Deep Dive: Click any launcher to see detailed performance over time including output views, copy/insertion, and feedback trends.

• High exposure, low interaction → launcher labels may be too vague or apply Microcopy Principles to improve clarity and specificity
• Low output copy/insertion rate → the generated content may not meet user needs or review and refine the launcher prompt
• High volume of end-user-created launchers → users are identifying gaps in builder-created actions or review popular custom actions and promote them
• Consistent negative feedback on a launcher → review the prompt, knowledge source, or output format for that specific action

• Total Users: Unique users shown an AI SmartTip suggestion (the user does not need to open the SmartTip bubble to be counted)
• Total Suggestions: Total number of AI SmartTip suggestions shown to users
• Total Successes: Number of times a user's input was accepted as valid by the AI
• Success Rate: Total successes ÷ total suggestions — the core measure of SmartTip effectiveness
• Topics: AI-generated groupings of SmartTips by theme, making it easier to analyse related field sets

• Suggestions vs. Success (trend graph): Tracks suggestion volume and success rate over time. A rising success rate indicates users are improving inputs in response to AI guidance, or that a prompt has been effectively refined.
• SmartTips Usage Trend: Time-series view of SmartTip exposure. Confirms SmartTips are reaching users after deployment and surfaces drop-off after system changes.
• SmartTips Overview (table): Per-SmartTip breakdown grouped by topic showing suggestions, successes, and success rate. Click any item to open the deep-dive side panel with performance over time.

• Low success rate on a specific SmartTip → the prompt may be too strict or poorly matched to how users write; refine the prompt language
• No improvement in success rate over time → check that user-facing feedback messages are specific enough to guide behavior change
• SmartTips with zero suggestions → check publishing status, segmentation rules, and whether the field is actively used
• Topics with consistently low success → escalate to the AI SME or the validation criteria may need a policy or process review

• Total Users: Unique users whose input triggered an AI rule evaluation
• Total Validations: Total number of times an AI rule evaluated a user input
• Triggered Rate: Proportion of evaluations where the condition was met and an action was taken
• True Validations: Total number of times the validation evaluated as true (condition met)
• False Validations: Total number of times the validation evaluated as false (condition not met)

• Users vs. Validations (trend graph): Tracks how often AI rules are evaluating inputs and how many users are affected over time.
• Rules Overview (table): All active AI rules with total validations and triggered rates. Click any rule to open a deep-dive side panel.
• AI Rule Deep Dive: A line graph comparing the percentage of unique users receiving true vs. false validations over time. Note: total validations may not equal plays for a piece of content, since the rule may be part of a larger display condition.

• Very high triggered rate → the condition may be too broad; narrow the prompt criteria in the Condition Builder
• Very low triggered rate → the condition may not be matching user inputs as expected or test with real examples and adjust the prompt
• Inconsistent triggered rate over time → check whether an underlying process or policy change has affected user input patterns

• Systems Overview: The main Insights landing page. Select any monitored system to access its dashboards, including the AI Dashboards tab.
• AI Dashboards tab: Available from the apps bar within any system overview. Displays up to four tabs based on your AI package: AI Answers, AI Action Bar, AI SmartTips, and AI Condition & Rule Engine.
• Custom Dashboards: Create personalized dashboards by adding metric widgets from different systems. Use to combine AI metrics with standard DAP metrics for stakeholder reporting.
• Reports & Report Builder: Build scheduled reports and subscribe stakeholders to automated delivery. Export widgets to CSV or XLSX for use in external reporting tools or executive presentations.
• Flow Analytics: How users move through workflows. Note: AI deployables are not available in Flow Analytics — use the AI Dashboards for AI-specific funnel data.